Introduction

This chapter considers the problem of estimation of a simple linear model with varying coefficients using alternative specifications for the varying coefficients. Thus to achieve the objective of obtaining estimators in the linear regression model people need to utilize some prior information about the coefficients _is and _is. Many different types of prior information are possible. These can be broadly classified as Systematic priors consisting of only nonstochastic terms; and stochastic priors consisting of both systematic and random components. In general the estimation of the systematically varying coefficients model is straightforward. The estimators obtained are both consistent and efficient. Another prior assumption regarding the varying coefficients in model could be that they vary systematically with the policy variable(s). The chapter indicates that there are alternative specifications for systematic variation of the coefficients in the model. It explains the estimation problem in various randomly varying coefficient models.